TALEN-induced gene knock out in Drosophila

Authors


Abstract

We report here a case study of TALEN-induced gene knock out of the trachealess gene of Drosophila. Two pairs of TALEN constructs caused targeted mutation in the germ line of 39% and 17% of injected animals, respectively. In the extreme case 100% of the progeny of TALEN-injected fly was mutated, suggesting that highly efficient biallelic germ line mutagenesis was achieved. The mutagenic efficiency of the TALEN pairs paralleled their activity of single strand annealing (SSA) assay in cultured cells. All mutations were deletion of 1 to 20 base pairs. Merit and demerit of TALEN-based gene knockout approach compared to other genome editing technologies is discussed.

Introduction

Manipulation of genes tagged with recombinant transposons with transposase and site-specific DNA recombinases permits mutagenesis and gene replacement of genomic loci (Rubin & Spradling 1982; O'Kane & Gehring 1987; Kawakami et al. 2000; Bischof et al. 2007; Venken et al. 2011). Although these technologies have considerably advanced molecular genetic studies in Drosophila, their application is still limited by incomplete coverage of transposons insertions. Designable site-specific DNA endonucleases have greatly expanded the opportunity of gene manipulation to untagged loci. The modular nature of nucleotide recognizing domains of zinc-finger type (Carroll 2011) or TAL effector type (Joung & Sander 2012) permits systematic design of specific DNA binding proteins. Pairs of such proteins recognizing nearby sequences of loci of interest are fused to Fok I DNA endonuclease. Properly designed pairs of such fusion proteins dimerize to form active nuclease and introduce double strand DNA breaks at specific genomic loci, and induce DNA recombination and deletion through non-homologous end joining (NHEJ). Although these strategies were well received, relatively few cases of actual use on Drosophila have been reported (Beumer et al. 2008; Katsuyama et al. 2013; Sakuma et al. 2013).

Here we applied the technique of TALEN on mutagenesis of the trachealess (trh) gene in Drosophila, which is located at the tip of the left arm of chromosome 3 (61C1). Although mutant alleles have been available, they are either missense mutation (Jürgens et al. 1984; Chung et al. 2011) or insertion/deletion at 5′ non-coding region (Isaac & Andrew 1996; Wilk et al. 1996), and no molecularly defined protein null mutation has been available. We report design and procedures of TALEN based gene knockout of trh and discuss the potential and the advantage of this method compared to other methods of gene mutagenesis in Drosophila.

Materials and methods

Construction of TAL effector nuclease

TAL effector DNA binding domain were designed and constructed based on modified ‘Golden Gate TALEN and TAL Effector Kit’ (Addgene) (Dahlem et al. 2012; Sakuma et al. 2013). DNA binding domain of each TALEN constructs are described as a chain of RVD repeat unit (NN, NI, HD, NG), “/” separates units of assembly.

trh A: Left, NN NN NI NN HD NG/NN HD NN HD NI NI/NN NN NI NN/NI; Right, HD HD NN HD NN HD/HD NN HD NN NI NG/HD NG HD/NN.

trh B: Left, HD NN NI NG HD NI/NG NI NI NN NI HD/NG NN NI/HD; Right, HD HD NN NN NI NN/NI NI NI NG HD HD/HD NG HD NI/NN.

trh C: Left, HD NG NI HD NN NI/NN HD NG NI NN HD/HD NI NI NN NI NG/NN/HD; Right, HD HD NI NN HD NG/NN NN HD NG NN NN/NG NN NI NG/NN.

trh D: Left, NN HD NI HD HD NG/NG NG NG NG HD NN/NG NG NN NI/HD; Right, HD NN NI HD NG NG/HD NG HD HD NG NG/NN HD NN HD NI NN/HD/NG.

trh E: Left, HD HD HD NG HD NG/NN HD HD NN NN HD/NI NN HD HD NI NG/HD/NI; Right, HD NG NG NI NG NN/NI NG HD NN NI NN/NN HD HD/NG.

Target sites for single strand annealing (SSA) assay were created by annealing complementary DNA strands for the following sequence with 5′-GTCG and 5′-CGGT extension for sense and antisense strand, respectively, and cloned into the BsaI site between the split luciferase fragments in pGL4-SSA vector (Sakuma et al. 2013).

A: GATGGAGCTGCGCAAGGAGAAGTCGAGGGAC-GCGGCGAGATCGCGGCGCGGAGGT.

B: GATCGATCATAAGACTGACCATCAGCTACCTG-AAGCTGAGGGATTTCTCCGGAGGT.

C: GATCGGGAAGCCTCTAGCAGCAGTAAGCTAAA-AAGTAAGTAACACAAATCTAGGT.

D: GATGCACCTTTTTCGTTGACTTTGCAGCATCCT-GGAGCTGCGCAAGGAGAAGTCGAGGT.

E: GATCCCTCTGCCGGCAGCCATCACCAGCCAG-CTGGACAAGGCCTCGATCATAAGAGGT.

Fly strains and microinjection

Synthetic mRNA was produced from linearized TALEN plasmids using the mMessage mMachine T7 Ultra Kit (Ambion, Austin, TX, USA) following the manufacturer's instructions and purified using the MEGAclear Kit (Ambion). mRNAs of TALEN pair A and B were mixed at a final concentration of 500 μg/mL in water and injected into the posterior end of 0–2 h y1 w1118 embryos using standard microinjection procedure. trh enhancer trap P{ET-L}trh-1-eve-1 (FBti0002897) (Perrimon et al. 1991) was also used as a host for TALEN pair A injection. Injected G0 flies were individually mated to y w; e Pr Dr/TM3, Sb flies and approximately 20 male progenies were tested for complementation of trh mutation by singly crossing to females of trh1 (FBal0017036) or trh3 (FBal0009624/FBal0063744) balanced over TM6B, Tb. New trh mutants were isolated from non-complementing crosses and rebalanced over TM6B, P{Ubi-GFP.S65T}PAD2, Tb1 (FBst0004887).

Genomic DNA sequencing

In order to sequence the targeted sites, three to five trh mutant 1st instar larvae were hand picked based on the lack of green fluorescent protein (GFP) fluorescence and air filled trachea. Larvae were homogenized in 10 μL of 10 mmol/L Tris–HCl, pH 8.5 with Nippi Biomasher II homogenizer (Nippi, Tokyo Japan) and the homogenates were used as a template for polymerase chain reaction (PCR) reaction using KOD FX Neo (TOYOBO, Osaka, Japan) and primers (Forward: TGCAGAGTCGTCTTGTGGAC, Reverse: CCTAAGAACAGTCTTATTGATGCT). Amplified DNA was purified using QIAquick Gel Extraction Kit (Qiagen, Hilden, Germany) and sequenced using one of the primers for the PCR reaction.

Results and discussion

Construction of TALEN pairs for trh

Trh protein contains conserved DNA binding bHLH domain and PAS-A and PAS-B domains. In order to disrupt Trh function, we designed to target TALEN cleavage site in the 236 base pair exon coding the bHLH domain common to all mRNA isoforms (Chung et al. 2011). This exon is spliced to the upstream exons containing the initiation codon and downstream exons coding PAS-A and PAS-B domains. Five pairs of TALEN were constructed and tested for target cleavage activity in the HEK293T cell using SSA assay (Fig. 1 A,B, Table 1) (Ochiai et al. 2010; Sakuma et al. 2013). In the pGL4-SSA reporter vector, the luciferase gene is split into two inactive fragments with 700-bp regions of homologous overlap separated by putative target cleavage sequences. Introduction of double strand DNA cleavage by cotransfected TALEN pair triggers a single strand annealing reaction between the homologous overlap regions, generating a functional luciferase gene. TALEN pair A and B induced significant luciferase activity and were chosen for mutagenesis assay.

Table 1. Summary of mutagenesis
TALEN pairHostFrequencyaG0trh/total
  1. a

    Number of G0 animals produced trh progeny/total number of G0.

trh Ay1 w11187/18A119/19
A108/12
A126/9
A133/10
A147/12
A164/5
A173/5
trh A1-eve-11/4A'19/11
trh By1 w11184/23B112/15
B161/19
B202/13
B231/16
Figure 1.

TALEN target sequences in the trh gene. (A) Sequence of trh exon 2. Adjacent intron sequences are indicated by lowercase. Five TALEN target sites (TALEN pair A–E) are indicated by color and underlines. (B) Single strand annealing (SSA) reporter assay of TALEN pairs. Each pair was tested for cleavage of respective target sequence by SSA assay in HEK293 cells. *For each pair, HPRT1-B reporter (Sakuma et al. 2013) was used as negative control with relative activity 1.0. **HPRT1-B TALEN pair (Sakuma et al. 2013) was included as positive control. (C) Mating scheme for isolation of new trh alleles.

mRNA injection and mutant selection

Synthetic mRNA was produced from linearized TALEN plasmids and mixtures of mRNAs of TALEN pair A and B (500 μg/mL each) were injected into the posterior end of 0–2 h y1 w1118 embryos. Surviving G0 flies were mated to strains with marked third chromosomes and up to 20 progenies of these crosses (G1) were mated to trh mutant strains to screen for new trh alleles (Fig. 1C). As shown in Table 1, 39% (7/18) of G0 flies injected with TALEN pair A yielded trh mutant. Percentage of trh mutants among G1 progenies varied from 30% to 100%. Injection of TALEN pair B yielded a slightly lower number of trh mutants: 17% (4/24) of G0 flies produced trh allele among 6% to 15% of G1 progenies. The difference in mutant induction between the two TALEN pairs paralleled the difference in their activity in the SSA assay. In addition, we injected TALEN pair A into trh-lacZ enhancer trap strain 1-eve-1 and obtained one G0 adult yielding trh mutants. The frequency of mutant induction was comparable to our previous study where GFP protein trapped gene jupitor was disrupted: 25% (7/28) of G0 produced 1.4% to 30.6% of G1 mutant progenies (Sakuma et al. 2013).

Molecular nature of trh mutations

trh alleles were sequenced from PCR amplified genomic DNA isolated from homozygous mutant embryos. As shown in Table 2 and 3, single G0 often produced multiple independent mutations. Deletions of 1 to 20 base pairs, but no insertion, were found within the region covered by the TALEN pairs. It should be noted that half of the sequenced alleles (14/28) were in-frame deletion (deletion of multiples of 3) that can retain the protein sequence downstream of mutagenized exon (PAS A and PAS B domain). This may permit a production of toxic protein product. One notable exception was mutant strain B11-1 in which we were unable to find any change in nucleotide sequence within the region franked by the recognition sequence of TALEN pair B. We speculate that double strand cleavage induced strand exchange between maternal and paternal chromosomes caused a mutagenic gene conversion event outside of the sequenced region. G0 founder A1 was special since all of its progeny was trh, suggesting that biallelic gene disruption occurred in its germ line. All mutants exhibited the phenotype of loss of trachea at the 1st instar larvae and larval lethality. Detailed characterization of mutant phenotype is under way and will be described elsewhere.

Table 2. Type of deletions induced by TALEN pair AThumbnail image of
Table 3. Type of deletions induced by TALEN pair BThumbnail image of

Efficiency of TALEN-mediated mutagenesis

Taking account of the number of G0 animals yielding no mutant progenies, frequency of the mutagenesis induced by TALEN pair A and B injected into y1 w1118 embryos was estimated to be 32% and 1.7% of G1 animals, respectively. These numbers are big enough for one to expect, with the frequency of TALEN pair B, to recover approximately eight mutant lines per 500 G1 progenies (sampling of ~20 progenies per G0). Thus, one can expect to isolate multiple mutants from 25 to 30 fertile G0 flies by unbiased DNA sequence analysis of PCR amplified target genomic sequence.

How TALEN compares to other mutagenesis method?

An ideal mutagenesis method must meet the requirements of simplicity and speed of procedure, expandability and high specificity. Here, we compare TALEN-mediated mutagenesis with other methods available. This and previous studies demonstrated that with properly designed TALEN pairs with appropriate activity in SSA assay, gene targeting can be performed in a relatively short period (~7 weeks). This is significantly shorter that the knock-in type gene targeting strategy involving the preparation of targeting vector and its integration into the fly genome (Rong & Golic 2000). Design and assembly of TAL effector modules is straightforward and permits quick assembly of full construct within 1 week: an advantage over a more complex procedure of zinc finger domains. The recently introduced CRISPR/Cas9 system provides a more versatile strategy to introduce DNA specificity to CAS9 DNA endonuclease protein by synthetic guide RNA (sgRNA) including 20 base pair homology to the target site (Jinek et al. 2012; Cong et al. 2013; Mali et al. 2013) and opened a flood of publications reporting successful gene disruption in a variety of organisms (Friedland et al. 2013; Wang et al. 2013) including Drosophila (Bassett et al. 2013; Gratz et al. 2013; Kondo & Ueda 2013; Yu et al. 2013). Clearly, introduction of specific sgRNA with common CAS9 protein is far simpler than the design and production of ZFN or TALEN. Therefore, in terms of simplicity, speed and expandability, the CRISPR/Cas9 system has an edge over other mentioned gene targeting methods.

Then what about the specificity? DNA binding proteins in general exhibit some degree of relaxed specificity. Cleavage specificity of mismatched TALEN targets has been studied in zebrafish. Dahlem et al. demonstrated that TALEN designed to ryr1b and ryr3 distinguished closely related gene family members (ryr1a, ryr1b ryr3) distinguished each others with mismatch of 6–7 base pair out of 36–37 base target sites, while mismatch of two (one perfect match by one of TALEN pair) permitted cleavage (Dahlem et al. 2012). Hisano et al. used two unique TALEN pairs and searched for potential cross-reactive genomic targets. Potential cleavage at those sites (3–6 base mismatch) was assessed after injection of TALEN mRNAs. While efficient cleavage was detected at the authentic sites, no evidence for cleavage at those potentially cross-reactive sites was detected (Hisano et al. 2013). Since closely resembling the target site can be avoided by proper design of TALEN, it should be possible to minimize the chance of background cleavage by TALEN. Specificity of the CAS9 system was also studied by assaying cross-reactivity of sgRNA against human globin beta to human globin delta (Cradick et al. 2013). Their results indicated that many constructs exhibited off-target cleavage of sites with 1–3 base mismatches out of 20 base recognition sites. Overall, while it is not possible to directly compare the chance of off-target effects, the CAS9 system that uses shorter 20 base target sequences has more risk of suffering from background cleavage.

Off-target cleavage and mutations can profoundly confound mutant analysis, which often takes a long time and effort to resolve. If one wants to characterize mutant phenotypes, it is desirable to start with clean loss of function alleles with minimal chance of accompanying mutation. In such a case investing a little more time for designing and testing several TALENs for best specificity may be a worthwhile investment.

Acknowledgments

This work was supported by MEXT/JSPS Grant-in-Aid for Young Scientists (B) Grant Number 24770197 (to TK), and Grant-in-Aid for Scientific Research on Innovative Areas Grant Number 22111007 (to SH) and Challenging Exploratory Research Grant Number 24657156 (to TY). TK is RIKEN Special Postdoctoral Fellow.

Ancillary